influence of molecular weight on the performance of organic solar cells based on a fluorene...

FULLPAPER

www.afm-journal.dewww.MaterialsViews.com

2124

Influence of Molecular Weight on the Performance ofOrganic Solar Cells Based on a Fluorene Derivative

By Christian Muller,* Ergang Wang, L. Mattias Andersson,

Kristofer Tvingstedt, Yi Zhou, Mats R. Andersson, and Olle Inganas

The performance of organic photovoltaic (OPV) bulk-heterojunction blends

comprising a liquid-crystalline fluorene derivative and a small-molecular

fullerene is found to increase asymptotically with the degree of polymerization

of the former. Similar to various thermodynamic transition temperatures as

well as the light absorbance of the fluorenemoiety, the photocurrent extracted

fromOPVdevices is found to strongly vary with increasing oligomer size up to

a number average molecular weight, Mn� 10 kgmol�1, but is rendered less

chain-length dependent for higher Mn as the fluorene derivative gradually

adopts polymeric behavior.

1. Introduction

Interest in fluorene-based organic semiconductors has surgedthanks to their ability to display high levels of (polarized) photo-and electroluminescence as well as useful electronic charge-transport properties.[1–3] More recently, a promising photovoltaicperformance has been achieved through incorporation of suchcompounds in bulk-heterojunction blends.[4–7]

Both, oligo- and polyfluorene derivatives have been studiedextensively and predominantly the latter have been found todisplay superior characteristics, such as significantly improvedopto-electronic behavior (e.g., field-effect mobility, photolumines-cence, circular dichroism, and photovoltaic performance),[8–11] aswell as, for instance, the ability to form oriented fibers.[12–14]

Generally, onewould expect thatwith increasing chain lengthmostproperties become less affected by small variations in the degree ofpolymerization, as indeed observed for the field-effect mobility,photoluminescence, as well as transition temperatures of variouspolyfluorenes.[9,14] However, frequently, it has been found that theviscosity of high molecular weight materials can limit theirprocessability, which in some cases had an adverse effect on,for example, the molecular ordering achievable with those

[*] Dr. C. Muller, Dr. L. M. Andersson, Dr. K. Tvingstedt, Dr. Y. Zhou,Prof. O. InganasDepartment of Physics, Chemistry & BiologyLinkoping University58183 Linkoping (Sweden)E-mail: [email protected]

Dr. E. Wang, Prof. M. R. AnderssonDepartment of Chemical and Biological EngineeringChalmers University of Technology41296 Goteborg (Sweden)

DOI: 10.1002/adfm.201000224

� 2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

species.[10,15,16] More advanced fluorenederivatives, such as some of those recentlydeveloped for organic photovoltaic (OPV)applications,[6,7] suffer, in addition, from thelimited solubility of higher molecular weightfractions in common organic solvents, whichfurther complicates their synthesis as well asprocessability. Thus, in order to permit anattractive (photovoltaic) performance whilstretaining good solution processability ofthese semiconductors, it may be necessaryto compromise between underperformingbut readily soluble oligomeric species andopto-electronically superior but less tractable

polymers.Here, we explore the influence of molecular weight on various

optical and physico-chemical properties of poly[2,7-(9,9-dioctyl-fluorene)-alt-5,5-(40,70-di-2-thienyl-20,10,30-benzothiadiazole)](commonly abbreviated as APFO-3 or PFDTBT,with ‘‘P’’ denotingpoly, but here as F8TBT in order to refer to both an oligo- andpolymeric material), as those can be readily related to the chainconformation. This we compare with the OPV performance ofblends with the small-molecular electron-acceptor [6,6]-phenyl-C61-butyric acid methyl ester (PC61BM). In particular, we find thatwith increasing molecular weight of F8TBT the photovoltaiccharacteristics improve and eventually become less sensitive tosmall changes in the degree of polymerization, offering a windowin terms of good processability as well as functionality.

2. Results and Discussion

Togain initial insight into themolecular conformationofF8TBT, ina first set of experiments we investigated the intrinsic viscosity, [h],of four representative batches, that is, 1, 3, 8, and 11 (see Table 1 formolecular weights), dissolved in ortho-dichlorobenzene (oDCB).Polymeric F8TBT should obey the Mark–Houwink–Sakuradaequation [h]¼K �Mn

a, which relates [h] and the number-averagemolecularweight,Mn,whereK anda are constants.[17]However, inthe log–log plot displayed in Figure 1, [h] doesn’t follow a simplelinear trend with respect to Mn but displays a concave tendency,suggesting that shorter, oligomeric species are of more rigidnature.[18] At higher Mn, the slope and hence the exponent aapproach �0.8, indicating that F8TBT molecules of sufficientchain length adopt a flexible to semi-flexible conformation inoDCB.

Adv. Funct. Mater. 2010, 20, 2124–2131

FULLPAPER

www.MaterialsViews.comwww.afm-journal.de

Table 1. Mn and Mw of F8TBT batches 1–14, relative to polystyrenestandards; polydispersity index, p.d.i.; degree of polymerization,d.p.¼Mn/M0, where M0¼ 686 gmol�1 is the molecular weight of aF8TBT repeat unit.

F8TBT batch Mn [kgmol�1] Mw [kgmol�1] p.d.i. d.p.

1 2 3 1.5 3

2 4 6 1.5 6

3 5 8 1.6 7

4 5 12 2.4 7

5 6 14 2.3 9

6 7 12 1.7 10

7 7 13 1.9 10

8 7 14 2.0 10

9 8 14 1.8 12

10 15 24 1.6 22

11 18 32 1.8 26

12 30 45 1.5 44

13 36 65 1.8 52

14 38 69 1.8 55

Figure 1. Chemical structure of F8TBT (top) and Mark–Houwink–Sakur-

ada plot of [h] with respect to Mn of 1, 3, 8, and 11 in oDCB (bottom).

Unfortunately, the limited availability of the remaining batches prevented

us from conducting a more complete study. A Mark–Houwink–Sakurada

exponent of a¼ 0.8 is indicated (generally, a� 0.5–0.8 for flexible poly-

mers; a> 0.8 for semi-flexible polymers; a¼ 2 if the species is a perfectly

stiff rod) [17]. Error bars represent the uncertainty in [h] as extrapolated

from dilution series.

Adv. Funct. Mater. 2010, 20, 2124–2131 � 2010 WILEY-VCH Verl

The semi-flexible character of F8TBT is corroborated by theappearance of liquid-crystalline textures in thin films that weresolidified by cooling from 300 8C to ambient as illustrated by aseries of polarized optical micrographs depicted in Figure 2.[7] It isworth noting that the observed domain size of such processedF8TBTdecreases considerably for 12–14, implying entanglementof polymer chains and, thus, we infer an entanglement molecularweight, Me� 18–30 kgmol�1.[19]

In order to further elucidate the transition of rigid short-chainoligomers to more random-coil semi-flexible polymers, wededuced the glass and liquid-crystalline/isotropic transitiontemperatures, Tg and Tlc/i, of 1–14 by means of differentialscanning calorimetry (DSC) and polarized optical microscopy(Fig. 3).Theheat stability ofF8TBTup to�320 8Cwasconfirmedbythermal gravimetric analysis (TGA) measurements (not shown).Conspicuously, a more ordered crystalline phase—readilyobserved for many other fluorene-based polymers[3,9,10,13–16]—is absent for the present material. The Flory–Fox equationTgðMnÞ ¼ TgðM1Þ þ B �M�1

n , where TgðM1Þ is the asymptoticvalue of Tg at infinite Mn and B a constant,[20] was employed todescribe the molecular weight dependence of Tg. For othermaterials systems a variety of characteristics, such as their tensilestrength, Tlc/i and the wavelength of maximum light absor-bance,[21–23] have been demonstrated to follow a similar trend,although more accurate models can be developed.[23] Throughoutthis study, a simpleM�1

n relationwas successfully used to illustratethe dependence of various properties onMn. It has been suggestedthat the Flory–Fox equation reflects a change in chain statistics dueto a decreasing influence of the chain-end free volume and thedevelopment of a random-coil conformation with increasingmolecular weight,[20,24] resulting in a largely chain-length inde-pendent behavior—particularly in absence of a crystalline phase.Accordingly, we find that the transition temperatures of F8TBTvary dramatically with molecular weight up to Mn� 10 kgmol�1.Thus, it appears reasonable to consider F8TBTas oligomeric belowand as more polymeric (macromolecular) above this threshold, asF8TBT gradually adopts a random-coil and eventually entangledchain configuration (i.e., for Mn>Me).

[24–27]

A change in chain conformation can be expected to drasticallyinfluence the optical properties of an organic semiconductor as itsdelocalized electronic transition states are related to the averageconjugation length along the backbone of the molecule, which, inaddition, may be constrained by the presence of chain ends, asdemonstrated for a variety of oligomer systems based on, forexample, paraphenylenevinylene[23,28] or 9,9-dihexylfluorene.[29]

F8TBT features two absorbance bands: a blue band below 400 nmattributed to a fully delocalized excitonic p–p� transition and a redband around 550 nm that has been ascribed to a localized charge-transfer state with the excited electron confined to the benzothia-diazole acceptor (B), whereas the corresponding hole remainsdelocalized across fluorene (F8) and thiophene (T) units along thep-conjugatedbackbone.[30]Delocalizationof these transition statesis prone to increase with oligomer size until the chain adopts asemi-flexible conformation. Hence, we find that both absorbancebands significantly red-shift with increasing oligomer molecularweight but saturate for higher degrees of polymerization (Fig. 4),confirming that F8TBT approaches macromolecular behavior forMn> 10 kgmol�1. [Note that a similar behavior has beenreported for thin films of poly[2-methoxy-5-(20-ethylhexyloxy)-

ag GmbH & Co. KGaA, Weinheim 2125

FULLPAPER


Figure 3. Tg (open circles) and Tlc/i, that is, Tlc!i and Ti!lc, (full circles)

versus Mn of 1–14 deduced from DSC second heating (top) and cooling

thermograms (bottom). Tlc/i were confirmed optically; 1 was found to be

monotropic as a liquid-crystalline phase was only observed during heating.

Trendlines were constructed assuming a M�1n relation (i.e., the Flory–Fox

equation for Tg).

Figure 2. Polarized optical micrographs of F8TBT films solidified by cool-

ing from 300 8C to ambient, that is, from T> Tlc/i to room temperature.

Note that 1 does not display a liquid-crystalline texture.

2126 � 2010 WILEY-VCH Verlag GmbH & C

1,4-phenylenevinylene] (MEH-PPV), for which, in addition,preferential in-plane orientation of chain segments was observedforhighmolecular-weightpolymers.[31]] Interestingly, the red-shiftin the absorbance maximum of F8TBTdissolved in oDCB and ofF8TBTsolidified in the presenceof an excessPC61BMfraction (i.e.,20:80 F8TBT:PC61BM) follows a comparable trend (Fig 4b),whereas the peak absorbance wavelengths, lp, of spin-coated andheat-treated F8TBT thin films saturate at a somewhat lowermolecular weight (Fig. 4c and d). Thus, the precise environmentappears to affect the chain flexibility of F8TBT (the value ofMn, forwhich lp saturates, can be taken as a measure for the effectiveconjugation length)[23] and in the case of theOPVblends discussedbelow the material can be considered as highly ‘‘diluted.’’

In order to relate the above observations to the OPVperformance of bulk-heterojunction blends comprising thepresent F8TBT oligomers and polymers, we fabricated a seriesof solar cells based on 1–14. The active layers of these devicescomprised 80wt% of the electron acceptor PC61BM—a composi-tion that has been reported to yield optimum power conver-sion[32]—and were spin-coated from oDCB.

o. KGaA, Weinheim Adv. Funct. Mater. 2010, 20, 2124–2131

FULLPAPER


Figure 4. a) Normalized UV-vis absorbance spectra of 1, 2, 4, and 13

dissolved in oDCB. b) Peak absorbance wavelengths, lp, versusMn of 1–14

in dilute oDCB solution as well as thin films of 20:80 F8TBT:PC61BM

solidified, that is, spin-coated, from oDCB [note that the blue peak below

400 nm was overshadowed by the absorbance of PC61BM]; c) thin films of

F8TBT spin-coated from oDCB; d) and thin films of F8TBT solidified by

cooling from 300 8C to ambient. Trendlines were constructed assuming a

M�1n relation.


Several reports have demonstrated that solidification of thismaterials combination from solutions with chlorinated benzenesyields a largely homogeneousmixture,[33–35] with phase separationlimited to partial crystallization of the small-molecular PC61BM,resulting in uniformly dispersed crystallites.[34,36] This is incontrast to solidification from other solvents such as chloroform,for which spinodal decomposition has been shown to precedevitrification of the blend, leading to a more inhomogeneousdistribution of F8TBT;[34] an effect thatmaybemore pronounced ifhigher molecular-weight polymers are employed. However,coarserF8TBT:PC61BMblends still feature significant intermixingon the molecular level, as evidenced by efficient photolumines-cence quenching reported in previous studies.[37,38]

Nevertheless, here, F8TBT can be considered to be ratherhomogeneously distributed in a matrix of the small-molecularPC61BM, corroborated by the smooth, featureless texture of thinfilms as revealed by a series of scanning force micrographs inFigure 5a. Despite the similar appearance of these blends, we findthat their OPV performance was significantly affected by theparticular choice of F8TBT batch (Fig. 5b). Whereas the open-

Figure 5. a) SFMs of 20:80 F8TBT:PC61BM thin films (z-range¼ 5 nm)

comprising 1 (top left), 2 (top right), 4 (bottom left), and 13 (bottom right).

b) Corresponding J–V characteristics of 20:80 F8TBT:PC61BM photovoltaic

devices based on 1, 2, 4, and 13 under 1000Wm�2 illumination.


FULLPAPER


2128

circuit voltage, Voc, was found to slightly decrease with increasingmolecularweight of F8TBT, especially the short-circuit current, Jsc,and to a limited extend also the fill factor, FF, were positivelyaffected, overall resulting in a much enhanced power-conversionefficiency,PCE (Fig. 6). Similar to various properties of F8TBT thathave been discussed above, such as Tg, Tlc/i, and the red-shift inabsorbance, the characteristics of F8TBT:PC61BM photovoltaicdevices are found to be proportional to M�1

n and to saturate forMn> 10 kgmol�1, indicating the influence of related microstruc-tural features, such as the number of F8TBT chain ends and thechain flexibility.

Figure 6. Average Voc, Jsc, FF, and PCE versusMn of 20:80 F8TBT:PC61BM

photovoltaic devices based on 1–14. Trendlines for Voc, Jsc, FF and PCE

were constructed assuming a M�1n relation. Error bars represent the

standard deviation of solar cell characteristics based on comparison of

similar devices.

� 2010 WILEY-VCH Verlag GmbH & C

The slight decrease inVoc can be rationalized by the red-shift inabsorbance of F8TBT with increasing oligomer size, which isindicativeof a change inenergy levelswith increasingchain length.As a result, the bandgap between the highest occupied molecularorbital (HOMO) of F8TBT and the lowest unoccupied molecularorbital (LUMO)ofPC61BMwill be reduced, leading to adecrease inphotovoltage with increasing molecular weight up toMn� 10 kgmol�1.[39,40]

Several factors may contribute towards the observed trend inphotocurrent generation. Certainly, an improved overlap betweenthe solar spectrum and F8TBT absorbance—red-shifting withincreasing oligomer size—will enhance the number of absorbedphotons,[11] virtually all of which will result in excitons thatdissociate into a weakly bound electron–hole charge pair (charge-transfer state) located at the interface of the electron-donating(hole-conducting) F8TBT and electron-accepting (electron-con-ducting) PC61BM.[41,42] However, a large fraction of charge pairsmay never contribute to the extracted photocurrent because ofexcessive recombination, which can be of geminate (monomole-cular) nature, referring to recombination of such a bound chargepair,[12,42–44] or of non-geminate (bimolecular) nature, referring tothe recombination of free charges (i.e., free electrons and holescreated by the complete dissociation of the bound charge pair) enroute to the electrodes.[45,46]

Undoubtedly, chargepairs generated at isolated sites (e.g., chainends)—a process that is more prevalent if short-chain oligomersare blended into the PC61BM matrix—will be strongly localized,which may prohibit efficient dissociation and, therefore, encou-rage their geminate recombination. In contrast, dissociation ofbound electron–hole pairs generated in proximity to longerpolymerchainsmaybenefit fromthemoreextendeddelocalizationof holes along the conjugated backbone (i.e., a higher intra-chainholemobility).[47] In addition, an increase in chain length and thusconnectivity of the hole-transporting F8TBT can improve inter-chain hole transport and, thus, may to some extent discouragebimolecular recombination of dissociated charges (a process thatcannot beneglected because of the insufficient phase separation ofthe present OPV blends).[48] Electron transport, however, may beless affected by changes in F8TBTmolecular weight as the small-molecular electron-conductor PC61BM can be expected toefficiently percolate throughout the active layer by virtue of thelarge excess in volume fraction. The benefit of higher F8TBTmolecular weight on hole transport is tentatively reflected in thesomewhat improved fill factor of F8TBT:PC61BM photovoltaicdevices andcorroboratedby the increase inblendholemobility,mh,thatwas extracted fromfield-effect transistor (FET)measurements(Fig. 7). Our FET devices did not display discernible electrontransport, which we assign to electron mobilities, me, that aresubstantially lower thanmh, as confirmed in Ref. [32] using photo-charge extraction by linearly increasing voltage (photo-CELIV), atechnique that for disordered systems can be correlated with FETmeasurements.[49] Unfortunately, we observed a pronouncedspread in FF and mh for devices produced with F8TBT of similarMn� 4–8 kgmol�1 (2–9), which to some extend may reflect yetunidentified differences in electronic purity between variousF8TBT batches. Interestingly, we find that, generally, a higher mh

results in an improved FF.Thus, for the present system, the above arguments suggest an

improvement in light absorbance and bound charge pair


FULLPAPER


Figure 7. Averagemh versusMn extracted from 20:80 F8TBT:PC61BM FETs

based on 1–14. Error bars represent the standard deviation of mh based on

comparison of similar devices.

Figure 8. Polarized optical micrograph (top; orientation of polarizer and

analyzer as indicated) and polarized photoluminescence spectrum (bot-

tom) of a F8TBT fiber produced with 14 (orientation of the polarizer with

respect to the fiber axis as indicated).

dissociation dynamics accompanied by reduced recombinationlosses as well as improved hole charge transport with increasingdegree of polymerization of F8TBT, resulting in significantlyenhanced photocurrent generation, which saturates forMn> 10 kgmol�1 as F8TBT assumes polymeric behavior. Here,it is interesting to note that the molecular-weight dependence ofthe OPV performance of similar blends composed of a relatedfluorene derivative, which comprises didecyl instead of dioctylside-chains (i.e., F10DTBT) and again PC61BM, was recentlyexplained principally on the basis of limited bound charge pairdissociation dynamics in low molecular-weight mixtures due torapid geminate recombination.[11] Besides, the same process wasalso found to be central for the photo-physics ofF8TBT(Mn� 5 kgmol�1):PC61BM blends.[42]

Furthermore, it is important to note that for the range of Mn

investigated in the present study, F8TBT with Me� 18–30 kgmol�1 is unlikely to be (strongly) entangled when dilutedwith PC61BM. However, chain entanglement for Mn>>Me canaffect the molecular order of the polymer,[10,15,16] and, hence, theOPV performance may eventually deviate from a simpleasymptotic trend. Similar behavior has been reported for blendscomprising the semicrystalline polymer poly(3-hexylthiophene)(P3HT) and PC61BM, which displayed a decrease in photovoltaicperformance when comprising P3HT of Mn> 34 kgmol�1,attributed to the negative effect of chain entanglement of highermolecular-weight P3HT on polymer crystallization.[50]

Finally, we would like to stress that any distinction between anoligomeric and polymeric regime is in fact strongly dependent ontheparticular property of interest. This ismost evidently illustratedby the polarized optical micrograph of a birefringent fiberdisplayed in Figure 8 (top), a macroscopic structure, which couldonly be produced with the highest available molecular-weightF8TBT batch (14). Such fibers featured pronounced orientation ofpolymer chains, as confirmed by polarized emission spectroscopy(Fig. 8, bottom). Clearly, although not suitable for certainfabrication schemes that rely on solubilization in more benignorganic solvents (c.f., Experimental), higher molecular weight


polymers open processing routes and permit structures that arenot readily accessible with oligomeric semiconductor species.

3. Conclusions

For the here discussed spin-coated bulk-heterojunction blendsbased on F8TBT and PC61BM, it appears sufficient to employrelatively low molecular-weight polymers with Mn> 10 kgmol�1

in order to guarantee good solution processability as well as asatisfactory OPV performance. We argue that ensuring thepolymeric nature of F8TBT is of benefit for light absorbance,charge dissociation as well as charge extraction, the relativeimportance of which is the subject of an ongoing investigation.Certainly, an asymptotic dependence on molecular weight can beexpected for othernon-crystalline polymer:fullereneOPVsystems.Moreover, materials selection based on a simple oligomer/polymer argument may also be relevant for other opto-electronicapplications comprising bulk-heterojunction blends, such as lowdark-current photo-detectors and light-emitting transistors.


FULLPAPER


2130

4. Experimental

Materials: F8TBT batches 1–14 were prepared according to previouslyreported procedures [6]; the molecular weight was controlled by varying thereaction time as well as by using an excess of the fluorenemonomer, which,therefore, can be expected to preferentially flank the synthesizedmaterial ateither chain end. The phenyl end-capped reaction product was carefullypurified by dissolution in chloroform (1–9) or oDCB (10–14) and additionof concentrated ammonia; after stirring for 24h the organic phase wasseparated and washed three times with distilled water, followed byprecipitation in methanol and finally collected by filtration. PC61BM wasobtained from Solenne BV.

Chlorinated benzenes—although not benign organic solvents—werechosen as solvents throughout this study as they readily permitteddissolution of all available batches at ambient; in contrast to chloroform ortoluene, for instance, which only dissolved lower molecular weight F8TBT.

Molecular Weight Determination: Number and weight average mole-cular weights, Mn and Mw, of 1–14 were determined using size exclusionchromatography (SEC) on a Waters 150 CV equipped with at refractiveindex detector using 1,2,4-trichlorobenzene as solvent at 135 8C and arelisted in Table 1. Calibration was performed with narrow molecular weightpolystyrene standards.

Viscometry: Intrinsic viscosities were determined using an Ubbelohdeviscometer from Schott (capillary number 0a). Dilution series withsolutions of 1, 3, 8, and 11 in oDCB (1–10 g dL�1) were recorded with aSchott AVS 360 controller.

Thermal Analysis: TGA was conducted under nitrogen at a scan rate of10 8Cmin�1 with a Perkin Elmer TGA 7 instrument. DSC was conductedunder nitrogen from 20 to 300 8C at a scan rate of 10 8Cmin�1 with a PerkinElmer Pyris 1 DSC instrument. Glass transition temperatures, Tg,correspond to inflection temperatures and liquid-crystalline/isotropictransition temperatures, Tlc/i, correspond to peak temperatures.

Thin Films And Fibers: Thin F8TBT films for microscopy and spectro-scopy studies were spin-coated on glass substrates from homogeneousoDCB solutions (�20 g L�1 F8TBT content) at ambient and, whereapplicable, thermally treated in a nitrogen-flushed Mettler FP82HT hotstage. Thin blend films (thickness �80–100 nm) for microscopy, spectro-scopy, and device studies were spin-coated from homogeneous oDCBsolutions (�30–60 g L�1 total material content; ratio of components20:80wt%:wt% F8TBT:PC61BM) at ambient. F8TBT fibers were drawn froma chunk of 14 after swelling in chloroform.

Optical Microscopy: Optical microscopy was carried out with anOlympus BH2 polarising microscope. Thermal transitions of F8TBT werestudied by heating/cooling thin films at a rate of 10 8Cmin�1 in a nitrogen-flushed Mettler FP82HT hot stage.

Scanning Force Microscopy (SFM): SFM was conducted with a VeecoDimension 3100 system in tapping mode using Antimony (n) dopedSilicon cantilevers (SCM-PIT, Veeco) with a force constant of 1–5Nm�1, aresonance frequency of 60–100 kHz and a tip curvature radius of 20 nm.

UV-vis Absorbance Spectroscopy: Polarized UV-vis absorbance spectra ofF8TBT and 20:80 F8TBT:PC61BM thin films as well as dilute oDCB solutions(0.1 g L�1; quartz cuvette with 1mm path length) were recorded using aPerkin Elmer Lambda 950 UV-vis spectrophotometer.

Photoluminescence Spectroscopy: A polarized photoluminescence emis-sion spectrum of a F8TBT fiber, exited at 525 nm, was recorded through alinear polarizing filter using an Oriel liquid light guide and a Shamrock SR303i spectrograph coupled to a Newton EMCCD silicon detector.

Photovoltaic Devices: Photovoltaic devices were fabricated onPEDOT:PSS coated (spin-coated from Baytron P, H. C. Stark GmbH,and treated at 150 8C for 30min, thickness �40 nm) patterned ITO-coatedglass substrates. Blend active layers (film thickness�80–100 nm; i.e., closeto a thickness of 85 nm, at which the power absorption of such devicesdisplays a local maximum [51]) were spin-coated as described above. A LiFelectron-blocking layer (thickness�6 A) and aluminum top electrodes weredeposited via thermal evaporation under vacuum (thickness �80 nm; area�4mm2).

Current density–voltage, J–V, characteristics were measured at ambientunder simulated solar illumination (Air Mass 1.5, 1000Wm�2). The light

� 2010 WILEY-VCH Verlag GmbH & C

source used was a 300 Watt xenon arc lamp solar simulator (PhotoEmission Tech.); the intensity was calibrated using a silicon photo-diodefrom Hamamatsu Photonics.

FETs: FETs were fabricated in bottom-gate bottom-contact configura-tion on highly doped silicon wafers with a thermally grown 100 nm siliconoxide layer; the two layers served as the gate electrode and gate insulator,respectively. Au source and drain electrodes, with a Cr adhesion layer weredefined by standard photolithography (channel length, L, varied from 9 to37mm; width,W, varied from 2 to 16mm to ensure a comparableW/L ratioand to exclude contact resistance effects). Blend active layers was spin-coated as described above.

Electronic characterization of FETs was conducted with a Keithley 4200parameter analyzer in high vacuum. Saturation field-effect hole mobilities,mh, were extracted from the transfer characteristics using

mh ¼@

ffiffiffiffiffiIsd

p

@Vg

� �22L

WCi(1)

where Isd is the source-drain current (saturation regime), Vg the gatevoltage, Ci the insulator capacitance, W and L the channel width andlength.

Acknowledgements

We acknowledge funding from the Northern European Innovative EnergyResearch Programme (N-INNER) as well as the Swedish Energy Agencythrough the projects Morphoso and Polarge.

Received: February 3, 2010

Revised: March 17, 2010

Published online: June 2, 2010

[1] Y. Ohmori, M. Uchida, K. Muro, K. Yoshino, Jpn. J. Appl. Phys. 1991, 30,

L1941.

[2] M. Grell, D. D. C. Bradley, M. Inbasekaran, E. P. Woo, Adv. Mater. 1997, 9,

798.

[3] H. Sirringhaus, R. J. Wilson, R. H. Friend, M. Inbasekaran, W. Wu,

E. P. Woo, M. Grell, D. D. C. Bradley, Appl. Phys. Lett. 2000, 77, 406.

[4] J. J. M. Halls, A. C. Arias, J. D. MacKenzie, W. Wu, M. Inbasekaran,

E. P. Woo, R. H. Friend, Adv. Mater. 2000, 12, 498.

[5] R. Pacios, D. D. C. Bradley, J. Nelson, C. J. Brabec, Synth. Met. 2003, 137,

1469.

[6] M. Svensson, F. Zhang, S. C. Veenstra, W. J. H. Verhees, J. C. Hummelen,

J. M. Kroon, O. Inganas, M. R. Andersson, Adv. Mater. 2003, 15, 988.

[7] O. Inganas, M. Svensson, F. Zhang, A. Gadisa, N. K. Persson, X. Wang,

M. R. Andersson, Appl. Phys. A 2004, 79, 31.

[8] D. J. Brennan, P. H. Townsend, D. M. Welsh, M. G. Dibbs, J. M. Shaw,

J. L. Miklovich, R. B. Boeke, A. C. Arias, L. Creswell, J. D. MacKenzie,

C. Ramsdale, A. Menon, H. Sirringhaus, Proc. SPIE 2003, 5217, 1.

[9] C. L. Donley, J. Zaumseil, J. W. Andreasen, M. M. Nielsen, H. Sirringhaus,

R. H. Friend, J.-S. Kim, J. Am. Chem. Soc. 2005, 127, 12890.

[10] R. Abbel, A. P. H. J. Schenning, E. W. Meijer, Macromolecules 2008, 41,

7497.

[11] D. J. D. Moet, M. Lenes, J. D. Kotlarski, S. C. Veenstra, J. Sweelssen,

M. M. Koetse, B. de Boer, P. W. M. Blom, Org. Electron. 2009, 10, 1275.

[12] M. Grell, D. D. C. Bradley, G. Ungar, J. Hill, K. S. Whitehead,Macromolecules

1999, 32, 5810.

[13] G. Lieser, M. Oda, T. Miteva, A. Meisel, H.-G. Nothofer, U. Scherf,

D. Neher, Macromolecules 2000, 33, 4490.

[14] M. Knaapila, R. Stepanyan, M. Torkkeli, B. P. Lyons, T. P. Ikonen, L. Almasy,

J. P. Foreman, R. Serimaa, R. Guntner, U. Scherf, A. P. Monkman, Phys. Rev.

E 2005, 71, 041802.

[15] M. J. Banach, R. H. Friend, H. Sirringhaus,Macromolecules 2003, 36, 2838.


FULLPAPER


[16] M. Knaapila, R. Stepanyan, B. P. Lyons, M. Torkkeli, T. P. A. Hase,

R. Serimaa, R. Guntner, O. H. Seeck, U. Scherf, A. P. Monkman, Macro-

molecules 2005, 38, 2744.

[17] L. H. Sperling, Introduction to Physical Polymer Science, 2nd ed., Wiley, New

York 2006; p 113f.

[18] M. K. Kosmas, A. M. Kosmas, Polymer 1993, 34, 3115.

[19] F. Elias, S. M. Clarke, R. Peck, E. M. Terentjev, Macromolecules 2000, 33,

2060.

[20] T. G. Fox, P. J. Flory, J. Appl. Phys. 1950, 21, 581.

[21] P. J. Flory, J. Am. Chem. Soc. 1945, 67, 2048.

[22] V. Percec, M. Kawasumi, Macromolecules 1993, 26, 3663.

[23] H. Meier, U. Stalmach, H. Kolshorn, Acta Polymer. 1997, 48, 379.

[24] Y. Ding, A. Kisliuk, A. P. Sokolov, Macromolecules 2004, 37, 161.

[25] J. M. G. Cowie, Eur. Polym. J. 1975, 11, 297.

[26] J. Hintermeyer, A. Herrmann, R. Kahlau, C. Goiceanu, E. A. Rossler,

Macromolecules 2008, 41, 9335.

[27] A. L. Agapov, A. P. Sokolov, Macromolecules 2009, 42, 2877.

[28] B. Tian, G. Zerbi, R. Schenk, K. Mullen, J. Chem. Phys. 1991, 95,

3191.

[29] D. Wasserberg, S. P. Dudek, S. C. J. Meskers, R. A. J. Janssen, Chem. Phys.

Lett. 2005, 411, 273.

[30] K. G. Jespersen, W. J. D. Beenken, Y. Zaushitsyn, A. Yartsev, M. Andersson,

T. Pullerits, V. Sundstrom, J. Chem. Phys. 2004, 121, 12613.

[31] K. Koynov, A. Bahtiar, T. Ahn, R. M. Cordeiro, H.-H. Horhold, C. Bubeck,

Macromolecules 2006, 39, 8692.

[32] L. M. Andersson, F. Zhang, O. Inganas, Appl. Phys. Lett. 2007, 91, 071108.

[33] F. Zhang, K. G. Jespersen, C. Bjorstrom, M. Svensson, M. R. Andersson,

V. Sundstrom, K. Magnusson, E.Moons, A. Yartsev, O. Inganas, Adv. Funct.

Mater. 2006, 16, 667.

[34] S. Nilsson, A. Bernasik, A. Budkowski, E. Moons,Macromolecules 2007, 40,

8291.


[35] C. M. Bjorstrom Svanstrom, J. Rysz, A. Bernasik, A. Budkowski, F. Zhang,

O. Inganas, M. R. Andersson, K. O. Magnusson, J. J. Benson-Smith,

J. Nelson, E. Moons, Adv. Mater. 2009, 21, 4398.

[36] D. Veldman, O. Ipek, S. C. J. Meskers, J. Sweelssen, M. M. Koetse,

S. C. Veenstra, J. M. Kroon, S. S. van Bavel, J. Loos, R. A. J. Janssen, J.

Am. Chem. Soc. 2008, 130, 7721.

[37] H. Aarnio, M. Westerling, R. Osterbacka, M. Svensson, M. R. Andersson,

H. Stubb, Chem. Phys. 2006, 321, 127.

[38] S. De, T. Kesti, M. Maiti, F. Zhang, O. Inganas, A. Yartsev, T. Pascher,

V. Sundstrom, Chem. Phys. 2008, 350, 14.

[39] A. Gadisa, M. Svensson, M. R. Andersson, O. Inganas, Appl. Phys. Lett.

2004, 84, 1609.

[40] M. C. Scharber, D. Muhlbacher, M. Koppe, P. Denk, C. Waldauf,

A. J. Heeger, C. J. Brabec, Adv. Mater. 2006, 18, 789.

[41] K. Vandewal, K. Tvingstedt, A. Gadisa, O. Inganas, J. V. Manca, Nat. Mater.

2009, 8, 904.

[42] S. De, T. Pascher, M. Maiti, K. G. Jespersen, T. Kesti, F. Zhang, O. Inganas,

A. Yartsev, V. Sundstrom, J. Am. Chem. Soc. 2007, 129, 8466.

[43] C. Groves, R. A. Marsh, N. C. Greenham, J. Chem. Phys. 2008, 129, 114903.

[44] M. Lenes, M. Morana, C. J. Brabec, P. W. M. Blom, Adv. Funct. Mater. 2009,

19, 1106.

[45] L. J. A. Koster, V. D. Mihailetchi, P. W. M. Blom, Appl. Phys. Lett. 2006, 88,

052104.

[46] C. G. Shuttle, B. O’Regan, A. M. Ballantyne, J. Nelson, D. D. C. Bradley,

J. R. Durrant, Phys. Rev. B 2008, 78, 113201.

[47] C. Deibel, T. Strobel, V. Dyakonov, Phys. Rev. Lett. 2009, 103, 036402.

[48] J. M. Frost, F. Cheynis, S. M. Tuladhar, J. Nelson, Nano Lett. 2006, 6, 1674.

[49] L. M. Andersson, F. Zhang, O. Inganas, Appl. Phys. Lett. 2006, 89, 142111.

[50] A. M. Ballantyne, L. Chen, J. Dane, T. Hammant, F. M. Braun, M. Heeney, W.

Duffy, I. McCulloch, D. D. C. Bradley, J. Nelson,Adv. Funct.Mater. 2008, 18, 2373.

[51] N.-K. Persson, H. Arwin, O. Inganas, J. Appl. Phys. 2005, 97, 034503.


influence of molecular weight on the performance of organic solar cells based on a fluorene...

Documents